you are pretty loose and free with your terminology.
there are optical mode locking where you pump a laser with one frequency, and expect another to come out...often helped by the laser resonator effect from the light bouncing off of semitransparent end mirrors. You would possibly control the
"color" of one of the lasers by fixing the beat note of the two optical sources to be a fixed frequency.
there are Optical delay lines where basically a microwave phase frequency locked loop is made by having a fiberoptic delay line that forms a very high gain frequency discriminator.
then you might actually be trying to phase lock an optical oscillator....like you use a delay line to impart a big phase shifter/frequency, modulate your microwave signal onto the optical delay line, demodulate your RF carrier some time delay later, pass it thru some sort of analog phase shifter, and a gain amplifier. If the gain is higher than unity, and the phase shift is 0 degrees (+/- maybe 25 degrees) it will indeed oscillate. basically you are making an standard transmission resonator oscillator, but using a fiberoptic delay line as the "resonator".
there are other ideas floating out there. there are four port optical devices where there are two inputs, one output proportional to phase shift, and a fourth output called a "vacuum port" that has coherent noise on it, and some signal processing can be done by it.
notice that only the THIRD option canbe a true Phase Locked Loop if you use a varactor tuned phase shifter.
all the others are "Frequenecy Locked" loops, that are not the same thing